Automatic blood glucose classification for gestational diabetes with feature selection: decision trees vs neural networks

Automatic blood glucose classification may help specialists to provide a better interpretation of blood glucose data, downloaded directly from patients glucose meter and will contribute in the development of decision support systems for gestational diabetes. This paper presents an automatic blood glucose classifier for gestational diabetes that compares 6 different feature selection methods for two machine learning algorithms: neural networks and decision trees. Three searching algorithms, Greedy, Best First and Genetic, were combined with two different evaluators, CSF and Wrapper, for the feature selection. The study has been made with 6080 blood glucose measurements from 25 patients. Decision trees with a feature set selected with the Wrapper evaluator and the Best first search algorithm obtained the best accuracy: 95.92%.

Brain structures identification based on feature descriptor

Traumatic Brain Injury -TBI- -1- is defined as an acute event that causes certain damage to areas of the brain. TBI may result in a significant impairment of an individuals physical, cognitive and psychosocial functioning. The main consequence of TBI is a dramatic change in the individuals daily life involving a profound disruption of the family, a loss of future income capacity and an increase of lifetime cost. One of the main challenges of TBI Neuroimaging is to develop robust automated image analysis methods to detect signatures of TBI, such as: hyper-intensity areas, changes in image contrast and in brain shape. The final goal of this research is to develop a method to identify the altered brain structures by automatically detecting landmarks on the image where signal changes and to provide comprehensive information to the clinician about them. These landmarks identify injured structures by co-registering the patient?s image with an atlas where landmarks have been previously detected. The research work has been initiated by identifying brain structures on healthy subjects to validate the proposed method. Later, this method will be used to identify modified structures on TBI imaging studies.

A scalable approach for content based image retrieval in cloud datacenter

The emergence of cloud datacenters enhances the capability of online data storage. Since massive data is stored in datacenters, it is necessary to effectively locate and access interest data in such a distributed system. However, traditional search techniques only allow users to search images over exact-match keywords through a centralized index. These techniques cannot satisfy the requirements of content based image retrieval (CBIR). In this paper, we propose a scalable image retrieval framework which can efficiently support content similarity search and semantic search in the distributed environment. Its key idea is to integrate image feature vectors into distributed hash tables (DHTs) by exploiting the property of locality sensitive hashing (LSH). Thus, images with similar content are most likely gathered into the same node without the knowledge of any global information. For searching semantically close images, the relevance feedback is adopted in our system to overcome the gap between low-level features and high-level features. We show that our approach yields high recall rate with good load balance and only requires a few number of hops.

Multi-Camera very wide baseline feature matching based on view-adaptive junction detection

This paper presents a strategy for solving the feature matching problem in calibrated very wide-baseline camera settings. In this kind of settings, perspective distortion, depth discontinuities and occlusion represent enormous challenges. The proposed strategy addresses them by using geometrical information, specifically by exploiting epipolar-constraints. As a result it provides a sparse number of reliable feature points for which 3D position is accurately recovered. Special features known as junctions are used for robust matching. In particular, a strategy for refinement of junction end-point matching is proposed which enhances usual junction-based approaches. This allows to compute cross-correlation between perfectly aligned plane patches in both images, thus yielding better matching results. Evaluation of experimental results proves the effectiveness of the proposed algorithm in very wide-baseline environments.

Signal analysis and feature generation for pattern identification of partial discharges in high-voltage equipment

This paper proposes a method for the identification of different partial discharges (PDs) sources through the analysis of a collection of PD signals acquired with a PD measurement system. This method, robust and sensitive enough to cope with noisy data and external interferences, combines the characterization of each signal from the collection, with a clustering procedure, the CLARA algorithm. Several features are proposed for the characterization of the signals, being the wavelet variances, the frequency estimated with the Prony method, and the energy, the most relevant for the performance of the clustering procedure. The result of the unsupervised classification is a set of clusters each containing those signals which are more similar to each other than to those in other clusters. The analysis of the classification results permits both the identification of different PD sources and the discrimination between original PD signals, reflections, noise and external interferences. The methods and graphical tools detailed in this paper have been coded and published as a contributed package of the R environment under a GNU/GPL license.

Relevance of the glottal pulse and the vocal tract in gender detection

Gender detection is a very important objective to improve efficiency in tasks as speech or speaker recognition, among others. Traditionally gender detection has been focused on fundamental frequency (f0) and cepstral features derived from voiced segments of speech. The methodology presented here consists in obtaining uncorrelated glottal and vocal tract components which are parameterized as mel-frequency coefficients. K-fold and cross-validation using QDA and GMM classifiers showed that better detection rates are reached when glottal source and vocal tract parameters are used in a gender-balanced database of running speech from 340 speakers.

Speech Signals Feature Extraction Model for a Speaker’s Gender and Age Identification System

Durante el proceso de producción de voz, los factores anatómicos, fisiológicos o psicosociales del individuo modifican los órganos resonadores, imprimiendo en la voz características particulares. Los sistemas ASR tratan de encontrar los matices característicos de una voz y asociarlos a un individuo o grupo. La edad y sexo de un hablante son factores intrínsecos que están presentes en la voz. Este trabajo intenta diferenciar esas características, aislarlas y usarlas para detectar el género y la edad de un hablante. Para dicho fin, se ha realizado el estudio y análisis de las características basadas en el pulso glótico y el tracto vocal, evitando usar técnicas clásicas (como pitch y sus derivados) debido a las restricciones propias de dichas técnicas. Los resultados finales de nuestro estudio alcanzan casi un 100% en reconocimiento de género mientras en la tarea de reconocimiento de edad el reconocimiento se encuentra alrededor del 80%. Parece ser que la voz queda afectada por el género del hablante y las hormonas, aunque no se aprecie en la audición. ABSTRACT Particular elements of the voice are printed during the speech production process and are related to anatomical and physiological factors of the phonatory system or psychosocial factors acquired by the speaker. ASR systems attempt to find those peculiar nuances of a voice and associate them to an individual or a group. Age and gender are inherent factors to the speaker which may be represented in voice. This work attempts to differentiate those characteristics, isolate them and use them to detect speaker’s gender and age. Features based on glottal pulse and vocal tract are studied and analyzed in order to achieve good results in both tasks. Classical methodologies (such as pitch and derivates) are avoided since the requirements of those techniques may be too restrictive. The final scores achieve almost 100% in gender recognition whereas in age recognition those scores are around 80%. Factors related to the gender and hormones seem to affect the voice although they are not audible.

Un modelo de relación entre los programas de innovación abierta y la creación de valor

En un mundo donde el cambio es constante y cada vez más vertiginoso, la innovación es el combustible que utilizan las empresas que permite su renovación constante y, como consecuencia, su supervivencia en el largo plazo. La innovación es sin dudas un elemento fundamental para determinar la capacidad de las empresas en crear valor a lo largo del tiempo, y por ello, las empresas suelen dedicar esfuerzos considerables y recursos de todo tipo para identificar nuevas alternativas de innovación que se adapten a su estrategia, cultura, objetivos y ambiciones corporativas. Una forma específica para llevar a cabo la innovación es la innovación abierta. Esta se entiende como la innovación que se realiza de manera conjunta con otras empresas o participantes del ecosistema. Cabe la aclaración que en este documento se toma la definición de ecosistema referida al conjunto de clientes, proveedores, competidores y otros participantes que interactúan en un mismo entorno donde existen posiciones de liderazgo que pueden cambiar a lo largo del tiempo (Moore 1996). El termino de innovación abierta fue acuñado por Henry Chesbrough hace algo mas de una década para referirse a esta forma particular de organizar la innovación corporativa. Como se observa en el presente trabajo la innovación abierta es un nuevo paradigma que ha capturado el interés académico y empresarial desde algo más de una década. Se verán varios casos de innovación abierta que se están llevando a cabo en diversos países y sectores de la economía. El objetivo principal de este trabajo de investigación es el de desarrollar y explicar un modelo de relación entre la innovación abierta y la creación de valor en las empresas. Para ello, y como objetivos secundarios, se ha investigado los elementos de un Programa de Innovación Abierta, los impulsores 1 de creación de valor, el proceso de creación de valor y, finalmente, la interacción entre estos tres elementos. Como producto final de la investigación se ha desarrollado un marco teórico general para establecer la conexión entre la innovación abierta y la creación de valor que facilita la explicación de la interacción entre ambos elementos. Se observa a partir de los casos de estudio que la innovación abierta puede abarcar todos los sectores de la economía, múltiples geografías y empresas de distintos tamaños (grandes empresas, pequeñas y medianas empresas, incluso empresas de reciente creación) cada una de ellas con distinta relevancia dentro del ecosistema en el que participan. Elementos de un Programa de Innovación Abierta La presente investigación comienza con la enumeración de los distintos elementos que se encuentran presentes en los Programas de Innovación Abierta. De esta manera, se describen los diversos elementos que se han identificado a través de la revisión de la literatura académica que se ha llevado a cabo. En función de una serie de características comunes, los distintos elementos se agrupan en cuatro niveles diferentes para lograr un mejor entendimiento de los Programas de Innovación Abierta. A continuación se detallan estos elementos § Organización del Programa. En primer lugar se menciona la existencia de una estructura organizativa capaz de cumplir una serie de objetivos establecidos previamente. Por su naturaleza de innovación abierta deberá existir cierto grado de interacción entre los distintos miembros que participen en el proceso de innovación. § Talento Interno. El talento interno asociado a los programas de innovación abierta juega un rol fundamental en la ejecución y éxito del programa. Bajo este nivel se asocian elementos como la cultura de innovación abierta y el liderazgo como mecanismo para entender uno de los elementos que explica el grado de adopción de innovación en una empresa. Estrechamente ligados al liderazgo se encuentran los comportamientos organizacionales como elementos diferenciadores para aumentar las posibilidades de creación de innovación abierta. § Infraestructura. En este nivel se agrupan los elementos relacionados con la infraestructura tecnológica necesaria para llevar a cabo el programa incluyendo los procesos productivos y las herramientas necesarias para la gestión cotidiana. § Instrumentos. Por último, se mencionan los instrumentos o vehículos que se utilizan en el entorno corporativo para implementar innovación abierta. Hay varios instrumentos disponibles como las incubadoras corporativas, los acuerdos de licenciamiento o las áreas de capital de riesgo corporativo. Para este último caso se hará una mención especial por el creciente y renovado interés que ha despertado tanto en el entorno académico como empresarial. Se ha identificado al capital de riesgo corporativo como un de los elementos diferenciales en el desarrollo de la estrategia de innovación abierta de las empresas ya que suele aportar credibilidad, capacidad y soporte tecnológico. Estos cuatro elementos, interactuando de manera conjunta y coordinada, tienen la capacidad de crear, potenciar e incluso desarrollar impulsores de creación de valor que impactan en la estrategia y organización de la empresa y partir de aquí en su desempeño financiero a lo largo del tiempo. Los Impulsores de Creación de Valor Luego de identificar, ordenar y describir los distintos elementos presentes en un Programa de Innovación Abierta se ha avanzado en la investigación con los impulsores de creación de valor. Estos pueden definirse como elementos que potencian o determinan la capacidad de crear valor dentro del entorno empresarial. Como se puede observar, se detallan estos impulsores como punto de interacción entre los elementos del programa y el proceso de creación de valor corporativo. A lo largo de la presente investigación se han identificado 6 impulsores de creación de valor presentes en un Programa de Innovación Abierta. § Nuevos Productos y Servicios. El impulsor de creación de valor más directo y evidente en un Programa de Innovación Abierta es la capacidad de crear nuevos productos y servicios dado que se relacionan directamente con el proceso de innovación de la empresa § Acceso a Mercados Adyacentes. El proceso de innovación también puede ser una fuente de valor al permitir que la empresa acceda a mercados cercanos a su negocio tradicional, es decir satisfaciendo nuevas necesidades de sus clientes existentes o de nuevos clientes en otro mercado. § Disponibilidad de Tecnologías. La disponibilidad de tecnologías es un impulsor en si mismo de la creación de valor. Estas pueden ser tanto complementarias como de apalancamiento de tecnologías ya existentes dentro de la empresa y que tengan la función de transformar parte de los componentes de la estrategia de la empresa. § Atracción del Talento Externo. La introducción de un Programa de Innovación Abierta en una empresa ofrece la oportunidad de interactuar con otras organizaciones del ecosistema y, por tanto, de atraer el talento externo. La movilidad del talento es una característica singular de la innovación abierta. § Participación en un Ecosistema Virtuoso. Se ha observado que las acciones realizadas en el entorno por cualquiera de los participantes también tendrán un claro impacto en la creación de valor para el resto de participantes por lo tanto la participación en un ecosistema virtuoso es un impulsor de creación de valor presente en la innovación abierta. § Tecnología “Dentro--‐Fuera”. Como último impulsor de valor es necesario comentar que la dirección que puede seguir la tecnología puede ser desde la empresa hacia el resto del ecosistema generando valor a partir de disponibilizar tecnologías que no son de utilidad interna para la empresa. Estos seis impulsores de creación de valor, presentes en los procesos de innovación corporativos, tienen la capacidad de influir en la estrategia y organización de la empresa aumentando su habilidad de crear valor. El Proceso de Creación de Valor en las Empresas Luego se ha investigado la práctica de la gestión basada en valor que sostiene la necesidad de alinear la estrategia corporativa y el diseño de la organización con el fin de obtener retornos financieros superiores al resto de los competidores de manera sostenida, y finalmente crear valor a lo largo del tiempo. Se describe como los impulsores de creación de valor influyen en la creación y fortalecimiento de las ventajas competitivas de la empresa impactando y alineando su estrategia y organización. Durante la investigación se ha identificado que las opciones reales pueden utilizarse como una herramienta para gestionar entornos de innovación abierta que, por definición, tienen altos niveles de incertidumbre. Las opciones reales aportan una capacidad para la toma de decisiones de forma modular y flexible que pueden aplicarse al entorno corporativo. Las opciones reales han sido particularmente diseñadas para entender, estructurar y gestionar entornos de múltiples incertidumbres y por ello tienen una amplia aplicación en los entornos de innovación. Se analizan los usos potenciales de las opciones reales como complemento a los distintos instrumentos identificados en los Programas de Innovación Abierta. La Interacción Entre los Programas de Innovación Abierta, los Impulsores de Creación de Valor y el Proceso de Creación de Valor A modo de conclusión del presente trabajo se puede mencionar que se ha desarrollado un marco general de creación de valor en el entorno de los Programas de Innovación Abierta. Este marco general incluye tres elementos fundamentales. En primer lugar describe los elementos que se encuentran presentes en los Programas de Innovación Abierta, en segundo lugar como estos programas colaboran en la creación de los seis impulsores de creación de valor que se han identificado y finalmente en tercer lugar como estos impulsores impactan sobre la estrategia y la organización de la empresa para dar lugar a la creación de valor de forma sostenida. A través de un Programa de Innovación Abierta, se pueden desarrollar los impulsores de valor para fortalecer la posición estratégica de la empresa y su capacidad de crear de valor. Es lo que denominamos el marco de referencia para la creación de valor en un Programa de Innovación Abierta. Se presentará la idea que los impulsores de creación de valor pueden colaborar en generar una estrategia óptima que permita alcanzar un desempeño financiero superior y lograr creación de valor de la empresa. En resumen, se ha desarrollado un modelo de relación que describe el proceso de creación de valor en la empresa a partir de los Programas de Innovación Abierta. Para ello, se han identificado los impulsores de creación de valor y se ha descripto la interacción entre los distintos elementos del modelo. ABSTRACT In a world of constant, accelerating change innovation is fuel for business. Year after year, innovation allows firms to renew and, therefore, advance their long--‐term survival. Undoubtedly, innovation is a key element for the firms’ ability to create value over time. Companies often devote considerable effort and diverse resources to identify innovation alternatives that could fit into their strategy, culture, corporate goals and ambitions. Open innovation refers to a specific approach to innovate by collaborating with other firms operating within the same business ecosystem.2 The term open innovation was pioneered by Henry Chesbrough more than a decade ago to refer to this particular mode of driving corporate innovation. Open innovation is a new paradigm that has attracted academic and business interest for over a decade. Several cases of open innovation from different countries and from different economic sectors are included and reviewed in this document. The main objective of this study is to explain and develop a relationship model between open innovation and value creation. To this end, and as secondary objectives, we have explored the elements of an Open Innovation Program, the drivers of value creation, the process of value creation and, finally, the interaction between these three elements. As a final product of the research we have developed a general theoretical framework for establishing the connection between open innovation and value creation that facilitates the explanation of the interaction between the two. From the case studies we see that open innovation can encompass all sectors of the economy, multiple geographies and varying businesses – large companies, SMEs, including (even) start--‐ups – each with a different relevance within the ecosystem in which they participate. Elements of an Open Innovation Program We begin by listing and describing below the items that can be found in an Open Innovation Program. Many of such items have been identified through the review of relevant academic literature. Furthermore, in order to achieve a better understanding of Open Innovation, we have classified those aspects into four different categories according to the features they share. § Program Organization. An organizational structure must exist with a degree of interaction between the different members involved in the innovation process. This structure must be able to meet a number of previously established objectives. § Internal Talent. Internal talent plays a key role in the implementation and success of any Open Innovation program. An open innovation culture and leadership skills are essential for adopting either radical or incremental innovation. In fact, leadership is closely linked to organizational behavior and it is essential to promote open innovation. § Infrastructure. This category groups the elements related to the technological infrastructure required to carry out the program, including production processes and daily management tools. § Instruments. Finally, we list the instruments or vehicles used in the corporate environment to implement open innovation. Several instruments are available, such as corporate incubators, licensing agreements or venture capital. There has been a growing and renewed interest in the latter, both in academia and business circles. The use of corporate venture capital to sustain the development of the open innovation strategy brings ability, credibility, and technological support to the process. The combination of elements from these four categories, interacting in a coordinated way, makes it possible to create, enhance and develop value creation drivers that may impact the company’s strategy and organization and affect its financial performance over time. The Drivers of Value Creation After identifying describing and categorizing the different elements present in an Open Innovation Program our research examines the drivers of value creation. These can be defined as elements that enhance or determine the ability to create value in the business environment. As can be seen, these drivers can act as interacting points between the elements of the program and the process of value creation. The study identifies six drivers of value creation that might be found in an Open Innovation Program. § New Products and Services. The more direct and obvious driver of value creation in any Open Innovation Program is the ability to create new products and services. This is directly related to the company’s innovation process. § Access to Adjacent Markets. The innovation process can also serve as a source of value by granting access to adjacent markets through satisfying new needs for existing customers or attracting new customers from other markets. § Availability of Technologies. The availability of technology is in itself a driver for value creation. New technologies can either be complementary and/or can leverage existing technologies within the firm. They can partly transform certain elements of the company’s strategy. § External Talent Strategy. Incorporating an Open Innovation Program offers the opportunity to interact with other organizations operating in the same ecosystem and can therefore attract external skilled resources. Talent mobility is a unique feature of open innovation. § Becoming Part of a Virtuous Circle. The actions carried out in the environment by any of its members will also have a clear impact on value creation for the other participants. Participation in a virtuous ecosystem is thus a driver for value creation in an open innovation strategy. § Inside--‐out Technology. Value creation may also evolve by allowing other firms in the ecosystem to incorporate internally developed under--‐utilized technologies into their own innovation processes. These six drivers that are present in the innovation process can influence the strategy and the organization of the company, increasing its ability to create value. The Value Creation Process Value--‐based management is the management approach that requires aligning the corporate strategy and the organizational design to create value and obtain sustained financial returns (at least, higher returns than its competitors). We describe how the drivers of value creation can enhance corporate advantages by aligning its strategy and organization. During this study, we were able to determine that real options can be used as managing tools in open innovation environments which, by definition, have high uncertainty levels. Real options provide capability for flexible and modular decision--‐making in the business environment. In particular, real options have been designed for uncertainty management and, therefore, they may be widely applied in innovation environments. We analyze potential uses of real options to supplement the various instruments identified in the Open Innovation programs. The Interaction Between Open Innovation Programs, Value Creation drivers and Value Creation Process As a result of this study, we have developed a general framework for value creation in Open Innovation Programs. This framework includes three key elements. We first described the elements that are present in Open Innovation Programs. Next, we showed how these programs can boost six drivers of value creation that have been identified. Finally, we analyzed how the drivers impact on the strategy and organization of the company in order to lead to the creation of sustainable value. Through an Open Innovation Program, value drivers can be developed to strengthen a company’s strategic position and its ability to create value. That is what we call the framework for value creation in the Open Innovation Program. Value drivers can collaborate in generating an optimal strategy that helps foster a superior financial performance and a sustained value creation process. In sum, we have developed a relationship model that describes the process of creating value in a firm with an Open Innovation Program. We have identified the drivers of value creation and described how the different elements of the model interact with each other.

A survey of feature selection in Internet traffic characterization

In the last decade, the research community has focused on new classification methods that rely on statistical characteristics of Internet traffic, instead of pre-viously popular port-number-based or payload-based methods, which are under even bigger constrictions. Some research works based on statistical characteristics generated large fea-ture sets of Internet traffic; however, nowadays it?s impossible to handle hun-dreds of features in big data scenarios, only leading to unacceptable processing time and misleading classification results due to redundant and correlative data. As a consequence, a feature selection procedure is essential in the process of Internet traffic characterization. In this paper a survey of feature selection methods is presented: feature selection frameworks are introduced, and differ-ent categories of methods are briefly explained and compared; several proposals on feature selection in Internet traffic characterization are shown; finally, future application of feature selection to a concrete project is proposed.

Expressive power of binary relevance and chain classifiers based on Bayesian Networks for multi-label classification

Bayesian network classifiers are widely used in machine learning because they intuitively represent causal relations. Multi-label classification problems require each instance to be assigned a subset of a defined set of h labels. This problem is equivalent to finding a multi-valued decision function that predicts a vector of h binary classes. In this paper we obtain the decision boundaries of two widely used Bayesian network approaches for building multi-label classifiers: Multi-label Bayesian network classifiers built using the binary relevance method and Bayesian network chain classifiers. We extend our previous single-label results to multi-label chain classifiers, and we prove that, as expected, chain classifiers provide a more expressive model than the binary relevance method.

Classifying GABAergic interneurons with semi-supervised projected model-based clustering

Objectives: A recently introduced pragmatic scheme promises to be a useful catalog of interneuron names.We sought to automatically classify digitally reconstructed interneuronal morphologies according tothis scheme. Simultaneously, we sought to discover possible subtypes of these types that might emergeduring automatic classification (clustering). We also investigated which morphometric properties weremost relevant for this classification.Materials and methods: A set of 118 digitally reconstructed interneuronal morphologies classified into thecommon basket (CB), horse-tail (HT), large basket (LB), and Martinotti (MA) interneuron types by 42 of theworld?s leading neuroscientists, quantified by five simple morphometric properties of the axon and fourof the dendrites. We labeled each neuron with the type most commonly assigned to it by the experts. Wethen removed this class information for each type separately, and applied semi-supervised clustering tothose cells (keeping the others? cluster membership fixed), to assess separation from other types and lookfor the formation of new groups (subtypes). We performed this same experiment unlabeling the cells oftwo types at a time, and of half the cells of a single type at a time. The clustering model is a finite mixtureof Gaussians which we adapted for the estimation of local (per-cluster) feature relevance. We performedthe described experiments on three different subsets of the data, formed according to how many expertsagreed on type membership: at least 18 experts (the full data set), at least 21 (73 neurons), and at least26 (47 neurons).Results: Interneurons with more reliable type labels were classified more accurately. We classified HTcells with 100% accuracy, MA cells with 73% accuracy, and CB and LB cells with 56% and 58% accuracy,respectively. We identified three subtypes of the MA type, one subtype of CB and LB types each, andno subtypes of HT (it was a single, homogeneous type). We got maximum (adapted) Silhouette widthand ARI values of 1, 0.83, 0.79, and 0.42, when unlabeling the HT, CB, LB, and MA types, respectively,confirming the quality of the formed cluster solutions. The subtypes identified when unlabeling a singletype also emerged when unlabeling two types at a time, confirming their validity. Axonal morphometricproperties were more relevant that dendritic ones, with the axonal polar histogram length in the [pi, 2pi) angle interval being particularly useful.Conclusions: The applied semi-supervised clustering method can accurately discriminate among CB, HT, LB, and MA interneuron types while discovering potential subtypes, and is therefore useful for neuronal classification. The discovery of potential subtypes suggests that some of these types are more heteroge-neous that previously thought. Finally, axonal variables seem to be more relevant than dendritic ones fordistinguishing among the CB, HT, LB, and MA interneuron types.

Enhancing regression models for complex systems using evolutionary techniques for feature engineering

This work proposes an automatic methodology for modeling complex systems. Our methodology is based on the combination of Grammatical Evolution and classical regression to obtain an optimal set of features that take part of a linear and convex model. This technique provides both Feature Engineering and Symbolic Regression in order to infer accurate models with no effort or designer's expertise requirements. As advanced Cloud services are becoming mainstream, the contribution of data centers in the overall power consumption of modern cities is growing dramatically. These facilities consume from 10 to 100 times more power per square foot than typical office buildings. Modeling the power consumption for these infrastructures is crucial to anticipate the effects of aggressive optimization policies, but accurate and fast power modeling is a complex challenge for high-end servers not yet satisfied by analytical approaches. For this case study, our methodology minimizes error in power prediction. This work has been tested using real Cloud applications resulting on an average error in power estimation of 3.98%. Our work improves the possibilities of deriving Cloud energy efficient policies in Cloud data centers being applicable to other computing environments with similar characteristics.

Finite element code-based modeling of a multi-feature isolation system and passive alleviation of possible inner pounding

The existing seismic isolation systems are based on well-known and accepted physical principles, but they are still having some functional drawbacks. As an attempt of improvement, the Roll-N-Cage (RNC) isolator has been recently proposed. It is designed to achieve a balance in controlling isolator displacement demands and structural accelerations. It provides in a single unit all the necessary functions of vertical rigid support, horizontal flexibility with enhanced stability, resistance to low service loads and minor vibration, and hysteretic energy dissipation characteristics. It is characterized by two unique features that are a self-braking (buffer) and a self-recentering mechanism. This paper presents an advanced representation of the main and unique features of the RNC isolator using an available finite element code called SAP2000. The validity of the obtained SAP2000 model is then checked using experimental, numerical and analytical results. Then, the paper investigates the merits and demerits of activating the built-in buffer mechanism on both structural pounding mitigation and isolation efficiency. The paper addresses the problem of passive alleviation of possible inner pounding within the RNC isolator, which may arise due to the activation of its self-braking mechanism under sever excitations such as near-fault earthquakes. The results show that the obtained finite element code-based model can closely match and accurately predict the overall behavior of the RNC isolator with effectively small errors. Moreover, the inherent buffer mechanism of the RNC isolator could mitigate or even eliminate direct structure-tostructure pounding under severe excitation considering limited septation gaps between adjacent structures. In addition, the increase of inherent hysteretic damping of the RNC isolator can efficiently limit its peak displacement together with the severity of the possibly developed inner pounding and, therefore, alleviate or even eliminate the possibly arising negative effects of the buffer mechanism on the overall RNC-isolated structural responses.

A study of feature combination for vehicle detection based on image processing

Video analytics play a critical role in most recent traffic monitoring and driver assistance systems. In this context, the correct detection and classification of surrounding vehicles through image analysis has been the focus of extensive research in the last years. Most of the pieces of work reported for image-based vehicle verification make use of supervised classification approaches and resort to techniques, such as histograms of oriented gradients (HOG), principal component analysis (PCA), and Gabor filters, among others. Unfortunately, existing approaches are lacking in two respects: first, comparison between methods using a common body of work has not been addressed; second, no study of the combination potentiality of popular features for vehicle classification has been reported. In this study the performance of the different techniques is first reviewed and compared using a common public database. Then, the combination capabilities of these techniques are explored and a methodology is presented for the fusion of classifiers built upon them, taking into account also the vehicle pose. The study unveils the limitations of single-feature based classification and makes clear that fusion of classifiers is highly beneficial for vehicle verification.

Nuevo algoritmo de compresión multimedia: codificación por saltos logaritmicos

Quizás el Código Morse, inventado en 1838 para su uso en la telegrafía, es uno de los primeros ejemplos de la utilización práctica de la compresión de datos [1], donde las letras más comunes del alfabeto son codificadas con códigos más cortos que las demás. A partir de 1940 y tras el desarrollo de la teoría de la información y la creación de los primeros ordenadores, la compresión de la información ha sido un reto constante y fundamental entre los campos de trabajo de investigadores de todo tipo. Cuanto mayor es nuestra comprensión sobre el significado de la información, mayor es nuestro éxito comprimiéndola. En el caso de la información multimedia, su naturaleza permite la compresión con pérdidas, alcanzando así cotas de compresión imposibles para los algoritmos sin pérdidas. Estos “recientes” algoritmos con pérdidas han estado mayoritariamente basados en transformación de la información al dominio de la frecuencia y en la eliminación de parte de la información en dicho dominio. Transformar al dominio de la frecuencia posee ventajas pero también involucra unos costes computacionales inevitables. Esta tesis presenta un nuevo algoritmo de compresión multimedia llamado “LHE” (Logarithmical Hopping Encoding) que no requiere transformación al dominio de la frecuencia, sino que trabaja en el dominio del espacio. Esto lo convierte en un algoritmo lineal de reducida complejidad computacional. Los resultados del algoritmo son prometedores, superando al estándar JPEG en calidad y velocidad. Para ello el algoritmo utiliza como base la respuesta fisiológica del ojo humano ante el estímulo luminoso. El ojo, al igual que el resto de los sentidos, responde al logaritmo de la señal de acuerdo a la ley de Weber. El algoritmo se compone de varias etapas. Una de ellas es la medición de la “Relevancia Perceptual”, una nueva métrica que nos va a permitir medir la relevancia que tiene la información en la mente del sujeto y en base a la misma, degradar en mayor o menor medida su contenido, a través de lo que he llamado “sub-muestreado elástico”. La etapa de sub-muestreado elástico constituye una nueva técnica sin precedentes en el tratamiento digital de imágenes. Permite tomar más o menos muestras en diferentes áreas de una imagen en función de su relevancia perceptual. En esta tesis se dan los primeros pasos para la elaboración de lo que puede llegar a ser un nuevo formato estándar de compresión multimedia (imagen, video y audio) libre de patentes y de alto rendimiento tanto en velocidad como en calidad. ABSTRACT The Morse code, invented in 1838 for use in telegraphy, is one of the first examples of the practical use of data compression [1], where the most common letters of the alphabet are coded shorter than the rest of codes. From 1940 and after the development of the theory of information and the creation of the first computers, compression of information has been a constant and fundamental challenge among any type of researchers. The greater our understanding of the meaning of information, the greater our success at compressing. In the case of multimedia information, its nature allows lossy compression, reaching impossible compression rates compared with lossless algorithms. These "recent" lossy algorithms have been mainly based on information transformation to frequency domain and elimination of some of the information in that domain. Transforming the frequency domain has advantages but also involves inevitable computational costs. This thesis introduces a new multimedia compression algorithm called "LHE" (logarithmical Hopping Encoding) that does not require transformation to frequency domain, but works in the space domain. This feature makes LHE a linear algorithm of reduced computational complexity. The results of the algorithm are promising, outperforming the JPEG standard in quality and speed. The basis of the algorithm is the physiological response of the human eye to the light stimulus. The eye, like other senses, responds to the logarithm of the signal according with Weber law. The algorithm consists of several stages. One is the measurement of "perceptual relevance," a new metric that will allow us to measure the relevance of information in the subject's mind and based on it; degrade accordingly their contents, through what I have called "elastic downsampling". Elastic downsampling stage is an unprecedented new technique in digital image processing. It lets take more or less samples in different areas of an image based on their perceptual relevance. This thesis introduces the first steps for the development of what may become a new standard multimedia compression format (image, video and audio) free of patents and high performance in both speed and quality.